Question

A data engineer defines a Lakeflow Declarative Pipeline (formerly Delta Live Tables) table with the following expectation: @dlt.expect("valid_amount", "amount > 0"). During a pipeline update, several records arrive with amount <= 0. What happens to those invalid records, and where can the engineer observe the violation?

Accepted Answer

The invalid records are retained in the target table, and the number of violations is reported in the pipeline's data quality metrics.

Answer

The invalid records are dropped from the target table, and the violation count appears in the event log.

Answer

The pipeline update immediately fails, and the invalid records are quarantined in a separate error table.

Answer

The invalid records are automatically corrected to a default value of 0, and no violation is logged.